Ascent–Drift–Descent Radiosonde System (ADDRS), China's independently developed third-generation radiosonde system, employs navigation satellite (Beidou and GPS) positioning technology to enhance the accuracy and reliability of wind measurements. Meanwhile, by using double sounding balloons nested inside and outside, it can conduct the three-phase observation of ascent–drift–descent, thereby achieving two vertical profiles and one stratospheric trajectory by releasing sounding balloons only once. Based on the data from the ADDRS GNSS (Global Navigation Satellite System) wind-finding radiosonde (hereafter ADDRS radiosonde) during the ascent and descent phases from field experiments conducted in Inner Mongolia and Guangdong Province, this study compared the observation-minus-analysis (O − A) statistics between the radiosonde data and the fifth generation European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis (ERA5) dataset. Furthermore, the data were assessed according to the uncertainty specified in the Observing Systems Capability Analysis and Review Tool (OSCAR) requirements, which are derived from the World Meteorological Organization (WMO) Rolling Review of Requirements (RRR) for upper-air measurements. Meanwhile, the operational L-band radar wind-finding radiosonde (hereafter L-band radiosonde) data at the same stations were taken as a reference. The results show that the wind components at all levels, as well as the temperature during nighttime, can be used directly. By contrast, it is recommended that positive bias corrections be applied to the temperature in the lower stratosphere in the ascent phase during daytime and the relative humidity at upper levels; while negative bias corrections be applied to the temperature at all levels in the descent phase during daytime and the relative humidity across the lower to middle troposphere, prior to distribution on the Global Telecommunication System (GTS) to increase their utility. Except for the relative humidity between 800 and 550 hPa, all other variables from the ADDRS radiosonde data perform better than the threshold values of OSCAR requirements uncertainty, demonstrating their suitability for numerical weather prediction (NWP) application. The wind components at all levels and the temperature below 100 hPa achieve the breakthrough values, indicating potential improvements for NWP performance.